Model Comparison
Medium 3
vs. Medium 3.5
Comparing 2 AI models · 12 benchmarks · Mistral
Recommended Pick
Mi Medium 3.5 11 metric wins Strongest on: Throughput, Reasoning, Intelligence
Best Value
Medium 3
100.0 value score
39.5 reasoning / $0.80/1M
Lowest Price
Medium 3
$0.40/1M input price
Best Reasoning
Medium 3.5
42.3 reasoning score
Blends available reasoning benchmarks
Best for Coding
Medium 3.5
35.4 coding index
Composite Indices
Higher is better; speed and price are normalized
Standard Benchmarks
Only benchmarks with data are shown
Differences That Matter
Best value
Medium 3 has the strongest quality-to-price mix at 100.0 out of 100 value points.
Price gap
Medium 3 is 3.8x cheaper on input tokens than Medium 3.5.
Speed gap
Medium 3.5 generates about 1.5x as many tokens per second as Medium 3.
Reasoning gap
Medium 3.5 leads Medium 3 by 2.8 points on reasoning.
Coding gap
Medium 3.5 leads Medium 3 by 21.8 points on coding.
Response Face-Off
Run one prompt through the selected models and compare response quality with live speed and cost context.
Medium 3
Mistral
TTFT
—
Time
—
tok/s
—
Tokens
—
Cost
—
Medium 3.5
Mistral
TTFT
—
Time
—
tok/s
—
Tokens
—
Cost
—
Which answer was more useful?
Full Comparison
| Metric | Mi Medium 3 | Top Pick Mi Medium 3.5 |
|---|---|---|
| Pricing per 1M tokens | ||
| Input Cost | $0.40/1M | $1.50/1M |
| Output Cost | $2.00/1M | $7.50/1M |
| Blended (3:1) | $0.80/1M | $3.00/1M |
| Specifications | ||
| Organization | Mistral | Mistral |
| Release Date | May 7, 2025 | Apr 29, 2026 |
| Performance & Speed | ||
| Throughput | 46.3 tok/s | 71.3 tok/s |
| TTFT | 555ms | 603ms |
| Latency | 555ms | 28635ms |
| Composite Indices | ||
| Value Score | 100.0 | 28.6 |
| Reasoning Score | 39.5 | 42.3 |
| Intelligence | 18.8 | 39.2 |
| Coding | 13.6 | 35.4 |
| Math | 30.3 | — |
| Standard Benchmarks | ||
| GPQA | 57.8% | 74.8% |
| MMLU Pro | 76.0% | — |
| HLE | 4.3% | 12.8% |
| LiveCodeBench | 40.0% | — |
| MATH 500 | 90.7% | — |
| AIME 2025 | 30.3% | — |
| AIME (Original) | 44.0% | — |
| SciCode | 33.1% | 39.6% |
| LCR | 28.0% | 61.0% |
| IFBench | 39.3% | 68.8% |
| TAU-bench v2 | 24.3% | 94.2% |
| TerminalBench Hard | 3.8% | 33.3% |
Key Takeaways
Medium 3 offers the best value at $0.40/1M, making it ideal for high-volume applications and cost-conscious projects.
Medium 3.5 has the strongest reasoning profile with a 42.3 reasoning score, combining the available reasoning-heavy benchmarks.
Medium 3.5 reaches a 35.4 coding index, making it the top choice for software development and code generation tasks.
All models support context windows of ∞+ tokens, suitable for processing lengthy documents and maintaining extended conversations.
When to Choose Each Model
Medium 3
- Cost-sensitive applications
- High-volume processing
Medium 3.5
- Complex reasoning tasks
- Research & analysis
- Code generation
- Software development
